netdev - Re: [PATCH net 1/2] net: ipv4: Set skb->dev for output route resolution

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite for Android: free password hash cracker in your pocket

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <CAF=yD-+qGQ42K2ebuY9YYHt260J--wqXqjWRqYXxP5XceVuz1g@mail.gmail.com>
Date:   Thu, 20 Dec 2018 20:51:08 -0500
From:   Willem de Bruijn <willemdebruijn.kernel@...il.com>
To:     Ido Schimmel <idosch@...lanox.com>
Cc:     "netdev@...r.kernel.org" <netdev@...r.kernel.org>,
        "davem@...emloft.net" <davem@...emloft.net>,
        "willemb@...gle.com" <willemb@...gle.com>,
        "dsahern@...il.com" <dsahern@...il.com>,
        Alexander Petrovskiy <alexpe@...lanox.com>
Subject: Re: [PATCH net 1/2] net: ipv4: Set skb->dev for output route resolution

On Thu, Dec 20, 2018 at 6:39 PM Ido Schimmel <idosch@...lanox.com> wrote:
>
> When user requests to resolve an output route, the kernel synthesizes
> an skb where the relevant parameters (e.g., source address) are set. The
> skb is then passed to ip_route_output_key_hash_rcu() which might call
> into the flow dissector in case a multipath route was hit and a nexthop
> needs to be selected based on the multipath hash.
>
> Since both 'skb->dev' and 'skb->sk' are not set, a warning is triggered
> in the flow dissector [1]. The warning is there to prevent codepaths
> from silently falling back to the standard flow dissector instead of the
> BPF one.
>
> Therefore, instead of removing the warning, set 'skb->dev' to the
> loopback device, as its not used for anything but resolving the correct
> namespace.
>
> [1]
> WARNING: CPU: 1 PID: 24819 at net/core/flow_dissector.c:764 __skb_flow_dissect+0x314/0x16b0
> ...
> RSP: 0018:ffffa0df41fdf650 EFLAGS: 00010246
> RAX: 0000000000000000 RBX: ffff8bcded232000 RCX: 0000000000000000
> RDX: ffffa0df41fdf7e0 RSI: ffffffff98e415a0 RDI: ffff8bcded232000
> RBP: ffffa0df41fdf760 R08: 0000000000000000 R09: 0000000000000000
> R10: ffffa0df41fdf7e8 R11: ffff8bcdf27a3000 R12: ffffffff98e415a0
> R13: ffffa0df41fdf7e0 R14: ffffffff98dd2980 R15: ffffa0df41fdf7e0
> FS:  00007f46f6897680(0000) GS:ffff8bcdf7a80000(0000) knlGS:0000000000000000
> CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
> CR2: 000055933e95f9a0 CR3: 000000021e636000 CR4: 00000000001006e0
> Call Trace:
>  fib_multipath_hash+0x28c/0x2d0
>  ? fib_multipath_hash+0x28c/0x2d0
>  fib_select_path+0x241/0x32f
>  ? __fib_lookup+0x6a/0xb0
>  ip_route_output_key_hash_rcu+0x650/0xa30
>  ? __alloc_skb+0x9b/0x1d0
>  inet_rtm_getroute+0x3f7/0xb80
>  ? __alloc_pages_nodemask+0x11c/0x2c0
>  rtnetlink_rcv_msg+0x1d9/0x2f0
>  ? rtnl_calcit.isra.24+0x120/0x120
>  netlink_rcv_skb+0x54/0x130
>  rtnetlink_rcv+0x15/0x20
>  netlink_unicast+0x20a/0x2c0
>  netlink_sendmsg+0x2d1/0x3d0
>  sock_sendmsg+0x39/0x50
>  ___sys_sendmsg+0x2a0/0x2f0
>  ? filemap_map_pages+0x16b/0x360
>  ? __handle_mm_fault+0x108e/0x13d0
>  __sys_sendmsg+0x63/0xa0
>  ? __sys_sendmsg+0x63/0xa0
>  __x64_sys_sendmsg+0x1f/0x30
>  do_syscall_64+0x5a/0x120
>  entry_SYSCALL_64_after_hwframe+0x44/0xa9
>
> Fixes: d0e13a1488ad ("flow_dissector: lookup netns by skb->sk if skb->dev is NULL")
> Signed-off-by: Ido Schimmel <idosch@...lanox.com>

Thanks a lot for addressing this warning, Ido!

David even already applied it, but still a belated

Acked-by: Willem de Bruijn <willemb@...gle.com>